UNCLASSIFIED

Defense Technical Information Center

Compilation Part Notice
ADP0I0421
TITLE: Recent Combat Aircraft Life Cycle Costing
Developments within DERA
DISTRIBUTION: Approved for public release, distribution unlimited

This paper is part of the following report:

TITLE: Design for Low Cost Operation and Support
[la Conception en vue d'une exploitation et d'un
soutien a cout reduit]
To order the complete compilation report, use: ADA388024
T~he component part is provided here to allow users access to individually authored sections
Af proceedings, annals, symposia, ect. However, the component should be considered within

the context of the overall compilation report and not as a stand-alone technical report.
T~he following component part numbers comprise the compilation report:
ADPOI0418 thru. ADP010432

UNCLASSIFIED

5-1

Recent Combat Aircraft Life Cycle Costing

Developments within DERA
Spencer Woodford
Centre for Defence Analysis, Lanchester Building, Defence Evaluation and Research Agency
Ively Road, Farnborough, GU14 OLX, United Kingdom.
ABSTRACT
In an effort to permit the procurement of more cost-effective
military equipment, several studies have been performed in
collaboration with two leading UK Universities. This paper
describes the rationale and requirements of both University
programmes, and gives details of the methods and some of the
results generated. Rather than a broad overview of many
different research activities within the Defence Evaluation and
Research Agency (DERA), the purpose of this paper is to give
as detailed a view as is possible of two recent studies, and the
future developments that will stem from them.
The first part of desin
the paper
tool developed
ado describes
comat aaftfor
iniumfor the
ife
doptimisation
Cycle Cost (LCC), whilst the second part explains the
evolution and optimisation of a long-range ground-attack
aircraft designed for minimum support. The LCC model
excludes 'deep overheads', restricting the use of the models to
the comparison of similar weapons systems (combat aircraft)
with a common set of design objectives and performance
constraints. The support estimation methodology of the
second part makes use of known aircraft design variables to
predict reliability and maintainability of the aircraft. Both
research activities, and the subsequent development at DERA,
should have a positive effect on the aircraft design process.
INTRODUCTION
The Defence Evaluation and Research Agency (DERA),
provides scientific advice, innovative engineering solutions,
of technical services to the UK Ministry of
range
aa broad
and
Defend
(boaD
r
The Centrehnicalsefnces
Analys
U Mie setory
o
Defence (MoD). The Centre for Defence Analysis is a sector
of DERA, and is primarily concerned with performing
operational analysis to provide authoritative and impartial
the Armed
advice to decision-makers within MoD and across
Forces. As part of a larger effort to reduce the cost of military
equipment, particularly operation and support cost, a number
of research studies were performed in collaboration with
Cranfield University and Imperial College, on the conceptual
design of combat aircraft for reduced support and LCC.
LCC is a complex subject that is concerned with quantifying
options to ascertain the optimum choice of assets and asset
configuration. When related to a combat aircraft, this leads to
the type of aircraft, its specification, and configuration. In
order to provide defensive and strike roles effectively in the
face of improvements in the potential enemies' forces, it has
been necessary to continually advance the performance,
capability, survivability, and support characteristics of the
aircraft, its associated weapon systems, and countermeasures,
This hns resulted in inrcrsinSg complexity nf sireraft and
systems and, in most instances, increasing costs, in both
absolute and real terms. Clearly, almost any new technology
could influence the LCC of the aircraft system. For this
reason this paper is not intended to be a comprehensive review
of all the LCC research taking place within DERA, but rather
an overview of two DERA-sponsored University research
programmes, and their intended development.

The US Department of Defence first applied Life Cycle

Costing to military projects in the early 1960's. It has become
more popular and important in the procurement of military
equipment, as the budgets for the World's fighting forces are
ever-increasingly tightened.
The reasons for this are
numerous and highly involved, needless to say that the end of
the Cold War, the global recession of the early nineteennineties, and the flood of low-initial-cost equipment from the
former Soviet Union have all played contributing roles.
With waning public support for defence expenditure, policy
makers must be seen to be cutting defence budgets in order to
facilitate increases in spending on welfare and other domestic
programs. Thus, military equipment must now be shown to
present 'value for money' in both the long and the short term.
'Value' is difficult to quantify in the military sense, leading
the current research activities to facilitate reductions in
through-life costs of aircraft designed for a specified level of
capability, mission performance, and operational requirement.
In this way, 'value' can be said to be maximised, as a set level
of performance is delivered for the lowest total cost.
In most previous studies of military aircraft, the objective
function (i.e. the variable subject of the optimisation) was
most often mass, either empty, mission, or gross mass. In the
civil world, direct operating cost is frequently the figure of
most interest to airlines, as it is the figure that allows the
operator to decide flight charges, and ultimately calculate
profit.
the militarycosts
environment,
'profit'
not just
shown,
although Inpeacetime
of operation
are isstill
as
a
just
il
ofoperatithere
osts
. It
al t
important. It would therefore appear that there is a need for a
greater understanding of the main contributors to the costs of
military equipment, not only for the acquisition phases, but
also in their operation and upkeep, and perhaps a re-think in
the way that equipment is designed.
The following sections of the paper briefly describe two
DERA-sponsored University research studies. The first is a
tool for the conceptual design of combat aircraft for minimum
LCC, and was performed by the author[l] whilst at Cranfield
University. The second describes a minimum support longrange, ground attack aircraft, the Low Support Vehicle (LSV),
and was performed by Whittle[2J, at Imperial College. There
are many similarities between the two pieces of work, but the
Low Support Vehicle was optimised for minimum mass, as no
discrete measure of support was determined.
The aircraft conceptual design tools used are based on
classical design methods, recently adapted and updated, and
validated with published data. The engine performance
modules consist of detailed thermodynamic models, modified
for the current usage. New engine sizing and mass estimation
routines were developed for both models. The LCC model is
primarily activity-based, and is an amalgamation of several
different methods, each written for a different phase in the
system life cycle. The LSV methodology makes use of two
measures of Support - the calculated levels of Reliability and
Maintainability (R&M) for the aircraft, and the number of
support aircraft required for a range of offensive missions.

Paper presented at the RTO A VT Specialists' Meeting on "Design for Low Cost Operation and Support",
held in Ottawa, Canada, 21-22 October 1999, and published in RTO MP-37.

5-2
PART I - CONCEPTUAL DESIGN OF COMBAT
AIRCRAFT FOR MINIMUM LIFE CYCLE COST
AIRCRAFT SYNTHESIS MODEL - The aircraft synthesis
and optimisation model is implemented via a large FORTRAN
code. Figure 1 gives a schematic representation of the overall
operation of the program. It can be seen that the optimiser has
ultimate control, and is responsible for altering the aircraft
design and engine sizing parameters such that all constraints
are met, and a minimum value of the objective function is
achieved.
For sufficiently accurate LCC prediction, the
synthesis model must have an appropriate level of fidelity, and
include realistic feature modelling and constraints. This was a
difficult balance to strike within the time constraints of the
study, and the LCC modelling routines may in future be added
to the more capable aircraft synthesis models developed
elsewhere within DERA.

is air mass flow rate, the value of which is determined by the

optimiser. The engine off-design analysis program is called at
sea level static conditions, and the values from this run,
together with the original engine design data, are used to
calculate the physical dimensions of the engine, using a
bespoke method. The engine intake area and maximum nozzle
area are generated for use in the aircraft geometry, mass, and
drag prediction methodologies.
The remainder of the aircraft is then sized so that an overall
configuration can be studied. The aircraft sizing process was
kept deliberately simple, in order to keep the number of
variables to a minimum, and improve robustness of the code.
A large number of design variables can cause the optimiser to
become trapped in local minima, and reduce the chances of
true convergence. Figure 2 shows the overall sizing of the
aircraft and the relevant major airframe design variables.
Intake Aea Oil

Input
Initialvaluesu

HFH

Arra

aSntshes
Ciii s

SgFN/
Inputt
inputuT

Input

i Omo ' .

Moel
,
.. ..
Engine

AircraftConfigonon
Perfornritcce
Req5

0-e

( ct isition)
perational D

Ign't
LimitseitBn

...
. ..................
......... .........
-X-

"..

..

Aircraft Synthesis
Models['.:
Life Cycle Const
(Objective Function)

------MPosEsMiissoon H A=:dyn:ni:

EngineModel

RDT&E
& IS
Opseraton
GiE

4 kSupport
V

Perfotoiance
Pins,
Miin

BW1F3
:Production
..

Dispraul

Constraint Analysis

.........
. ....................

Field Peffornuance, Attained Turn Sustained Turn.

Mach Number.Accaleration/Climb, Specific Excems
__Power.F.uel._Man,.FuelV~olum.
e, D~irr_
nion _Clks

Figure 1. Overall Program Operation.

The first stage in the aircraft synthesis procedure is to read the
relevant input data files and set the required parameters. The
design options available include the aircraft type,
configuration, number of crew, number of engines, etc.
Further input data specifies the overall design requirements of
the aircraft, including maximum level Mach number, diving
Mach number, limit load factor, maximum payload, avionics
mass, the number of weapon pylons, and other design drivers,
The parameter initialisation process also calls the engine
design program. The engine thermodynamic cycle design is
performed using data from the engine input file, which also
contains off-design limits for the engine. This adds to the
realism of the model by restricting the engine operating
envelope. Once the cycle of the engine has been set it is not
altered, and all subsequent engine calculations are performed
to analyse the engine performance away from its design point,
The above actions (i.e. file read and engine thermodynamic
design) are only performed on the very first call to the
synthesis.
Ali of the following design procedures are
performed every time the synthesis is called by the optimiser.
Component Sizing - Although the engine thermodynamic
cycle has been specified, the physical size of the engine is yet
to be defined. The main parameter used to determine engine
size (in terms of both engine thrust and physical dimensions)

Figure 2. Aircraft Variable Definitions.

The size of the fuselage is determined using the maximum
fuselage length, and the maximum effective fuselage diameter.
From these two variables, and engine parameters calculated
earlier, the remaining fuselage dimensions can be estimated.
The width and height dimensions are driven by the maximum
effective diameter, whether the aircraft has one or two engines,
and the size of the engine(s). Both the height and width could
have been varied separately, but a single variable was felt to
be beneficial, as explained above. For both single and twinengine aircraft, constraints are added to ensure that the
fuselage cross-sectional area is large enough to accommodate
the engine(s), and that the height is at least 20% larger than
the maximum diameter of the engine.
The optimiser provides values for gross wing area, aspect
ratio, taper ratio, leading-edge sweep, and thickness/chord
ratio; all other wing variables, including tip and centreline
chords, are found from standard geometry calculations. These
values are used in the calculation of aerodynamic performance
and wing fuel storage volume. The sizing of the empennage is
performned using parametric sizing equations developed for
this methodology, and the results are used in the aircraft mass
and drag estimation procedures. For both the tail and fin,
other parameters are also calculated, namely aspect ratio,
thickness/chord ratio, and mean chord. All wing, fin, and tail
parameters required by the synthesis have now been
generated, concluding the geometric definition of the aircraft.

5-3
Mass Estimation and Volume Accountine - One of the most
important processes in the design of any aircraft is the
estimation of the aircraft mass, which in this methodology, is
calculated from the sum of the individual component masses.
However, many of the component masses are themselves a
power function of the aircraft all-up mass, and the process
becomes an iterative procedure to converge on the correct
mass of the current design configuration.

performance characteristics. ONX performs this section of the

engine design, and from relatively simple starting values, the
nature of the engine cycle is determined. The user can
determine the design cycle of the engine simply by changing
the parameters in the input file.
Once the engine cycle and limitations have been set, the
engine is analysed away from the design point by the offdesign analysis program, OFFx. This program returns all of

The mass estimation method is implemented in such a way as

to mimic the historical use of composite materials; the first
structural component mass estimated is that for the
empennage, followed by the wing, and finally the fuselage.
Systems masses are found using semi-empirical methods, with
separate parametric equations for each of the major systems.
Fuel mass is calculated based on a fuel fraction value supplied
by the optimiser. From the mass of structure, systems, and
fuel, the aircraft gross mass and mission masses are calculated.
Although the above methods are relatively simple, Figure 3
shows that surprisingly accurate results are achieved when the
aircraft is treated as a whole system.

the major performance parameters for a particular engine and

flight condition. From these values, and calculating the
installation losses, the thrust available and fuel bum can be
estimated at any flight condition and throttle setting. An
optimiser constraint has been added to ensure that the thrust
available meets or exceeds the required thrust at all flight
conditions. The constraint is very useful, as it allows the
thrust at every mission phase to be checked using only a single
variable. It ensures that the aircraft can complete supercruise
and other high-thrust mission legs without the need for extra
point performance constraints, for which the aircraft mass will
not be known at the mission definition phase.

2=

.have
powerplant

-------------.

114
X

LW

.......

R=

F46

Once the engine air mass flow rate has been established for a
particular application and the design choices and limitations
been set, the mass and physical dimensions of the
are calculated. Continuing a theme suggested by
Whittle[2], a new engine dimension and mass estimation
methodology has been developed. The new models are based
on the major engine design drivers; air mass flow rate, bypass
ratio, compressor pressure ratio, number of shafts (although

"ONx &

OFrx only deal with two-shaft engines), and reheat

increase. The results are very promising, but there is a
question to be resolved over the accuracy of engine mass

r4 .thrust
F.
0

(-dnmpyM-ft

Figure 3. Mass Estimation Correlation.

The final process in this section of the code is to ensure that
there is sufficient volume available for fuel carriage. Assumed
system densities are used to subtract relevant volumes from
the total available in the wings and the fuselage, and is
implemented as an optimiser constraint,
Aerodynamic Modelling - The aerodynamics module consists
of three models. The first predicts available lift coefficient
based on wing configuration and geometry, Mach number, and
the presence of high-lift devices. The second calculates the
angle of attack from the lift-curve slope, which is based on the
clean wing geometry and flight Mach number, and contains a
simple correction for the effects of vortex lift. The third
section is the largest and most complex of the three, and
calculates the drag of the aircraft based on its geometry, lift
coefficient, configuration, and the presence of external stores
and retractable components. Due to the level of complexity of
the models, and in the interests of brevity, the aerodynamic
models are not expanded further in this paper.
Propulsion Modelling - Propulsion modelling is performed
using two thermodynamic codes, ONX and OFFx, written by
Mattingley[3]. The capability of the models has been limited,
for the purposes of this study, to reheated turbojets and
turbotans, and several improvements have been made from the
original codes. Engine design starts with on-design analysis,
which presumes that all design choices are still under control
and that the size of the engine is yet to be fixed. The
performance parameters are given as 'specific' values,
normalised with engine size, and each complete set of design
choices will result in an engine with its own operating and

prediction in the 125-175kg/s air mass flow rate range.

Point and Mission Performance

Point performance

calculations are used to compare the delivered performance of

the designed aircraft with the required performance figures.
They play a crucial role in the sizing of the aircraft, as the
performance constraints determine the aircraft wing loading
and thrust/weight ratio. The sizes of the wing and engine have
a major impact on the overall design of the aircraft, and
therefore the point performance calculations must be accurate,
if a realistic design is to be produced. The synthesis is able to
consider up to ten different point performance constraints; the
amount of fuel, payload, engine operation (maximum or
military thrust), and the individual point performance level
can be specified. Of the ten available point performance
constraints, the user has a choice of seven different constraint
types. These include take-off and landing, attained turn rate,
sustained turn rate (both in either g or */s), specific excess
power, maximum speed, and time-to-climb/acceleration.
Maximum height can also be calculated, but is not included as
a constraint.
The mission performance calculations work, for the most part,
in a similar manner to the point performance constraint
analysis methods, many of the algorithms being identical. The
main difference in this section is that the major factor being
calculated is the amount of fuel burned for each mission leg.
The sum of all of these masses, plus a user-defined reserve
factor, gives the total mission fuel mass, one of the single most
important values in the sizing of the aircraft. Up to thirty
mission legs can be specified from eight phase types, those
being; engine run, take-off, climb/accelerate/descend, cruise,
combat manoeuvres, weapons drop, loiter/CAP, and landing.
Range credit is ignored for climb/accelerate/descend and loiter
phases. Supercruise legs are specified by setting the cruise
Mach number, and restricting the use of reheat.

5-4

LIFE CYCLE COST MODEL - The LCC module is based on

several models that have been acquired and developed from
many different sources, and has been split into the areas most
often quoted in the available literature. Those are Research,
Development, Test and Evaluation (RDT&E); Production;
Ground Support Equipment and Initial Spares (GSE&IS);
Operation and Support (O&S), and Disposal. Each life cycle
phase model is represented by a subroutine, with all of the
data coming either from the aircraft synthesis models, or the
LCC input file. This file contains data such as procurement
and operation data, production rates, fuel costs, as well as cost
factors for security, flight test, and stealth considerations.

Production costs typically account for 30-35% of the total

LCC of modem combat aircraft; the cost per aircraft decreases
with number of aircraft built, as 'learning curve' theory and
economies of scale are applied. The airframe production cost
models are taken from the activity-based cost model derived
by Bums, whilst the engine cost model is taken from a USAF
Flight Dynamics Laboratory report[6]. Avionics cost models
are based on uninstalled avionics mass. The total of the two
major costs above (RDT&E & Production), divided by the
total number of aircraft built, is called the Unit Acquisition
Cost.
Ground Support Equipment and Initial Spares - This area of

Research, Development, Test, and Evaluation - The RDT&E

phase covers all areas of research and development prior to
full-scale production of the first production aircraft. It
includes; concept definition, design studies and integration,
wind tunnel models and testing, laboratory testing, production
of static and flight test airframes, avionics, software,
propulsion development, flight testing, integrated logistics
support, and program oversight. RDT&E typically makes up
about 10-15% of the LCC of modern, low-production ( 500)
abou
te LC
1015%
ofmodm,
f
lw-podution(- 00)
combat aircraft, but is obviously affected by the number of
aircraft over which this cost can be amortised. The method for
the calculation of airframe development costs is taken directly
from a method developed by Burns[4].

LCC is very difficult to estimate because of the equipment

requirements for a particular weapon system. In keeping with
suggestions made by several members of the costing
community, GSE&IS cost is simply a fraction of the aircraft
recurring flyaway cost, resulting in about 5% of total LCC.
Operation and Suppo - Operation and support (O&S) costs
for moder
combat aircraft can be split into several parts, all
f modem
co
ntrirtrsft
thest
o singral
mbat
of which will be contributors to the cost of using combat
aircraft in a peacetime operating regime. In wartime, the cost
aircratin apatmort
regime. In wartime,
th
of opraes
muche
portantwith
all reso
made operation
av ion personnel;
the support
particlrconfli
O&S
costs scomprise:
personnel;
service allowances, personnel support, and training; unit level

The methodology breaks the development procedure into

many different activities, with the effort for each being
estimated, and then multiplied by an appropriate labour rate to
calculate cost. It can thus be thought of as an 'Activity-Based
Costing' (ABC) procedure. The airframe cost model is based
on parametric estimating techniques, using airframe mass, and
several user-specified or design-dependent factors to allow for
differences in the designs. The method used for engine
development cost estimation is taken from a model developed
by Birkler[5]. It uses thrust, Mach number, and turbine inlet
temperature as the main cost drivers, and has been found to be
accurate for the limited data available. Avionics and software
development costs have proven problematic; avionics cost is
based on uninstalled mass, whilst software cost estimation
uses the number of lines of code and a number of user-defined
complexity factors.

consumption; contract costs for airframe, avionics, propulsion,

and supply; sustaining support funds; and basing overheads
and upkeep. The breakdown of O&S costs used follows the
methods and structure suggested by the US Office of the
Secretary of Defence in their 'Operating and Support Cost
Estimating Guide'[7]. The O&S cost is calculated for one
Main Operating Base, as this will have a significant effect on
the staff requirements for the particular aircraft. This gives the
added benefit that the model allows different basing concepts
to be investigated, whereas many previous cost models
considered the total number of aircraft procured.

Production - This includes production engineering design,

production investment (manufacturing facilities, tooling, jigs
and fixtures), manufacturing labour, quality control, material
and equipment, profit, overheads, administration, and
purchasing of engines and avionics systems. These costs,
divided by the number of aircraft make up the Recurring
Flyaway Cost. Figure 4 shows a comparison of quoted and
calculated recurring flyaway costs.

S $70y

=0.99,7x

.=0.9937
$60

*F-14

k7
.7
F:r-14

..-

rS40

19p
p-

S~~~~~~~F-16C

so

so

H~f

GR"

.airframe,
$10

340
$50
560
Quoted Recurring Flyaway Cost (1996$M)

$20 . S30

570

Figure 4. Recurring Flyaway Cost Comparison (FY96).

$80

One of the largest single O&S costs is that for mission

vast majority of whom are involved with First
personnel,
and SecondtheLine
maintenance. In order to calculate the
n the to fiscsiates
p
maintenance
number of
number of maintenance personnel, the method first estimates
the total maintenance effort required by the aircraft, using a
parametric method. The number of First and Second Line
operation personnel is calculated from the number of aircraft,
crew ratio, the annual flying time per aircraft, and the total
maintenance effort per flying hour. Support personnel
numbers are calculated from the number of operations
personnel and the number of aircraft. A separate section of
the O&S model estimates the costs of Officer and Enlisted
personnel training costs, training funds, and permanent change
of station allowances. The models were adapted and updated
from a US Navy report[8], and the values were calibrated
against RAF Cost Of Support Spreadsheet data.
Unit Level Consumption attempts to capture the costs for all
consumables used in operating the aircraft, including fuel, oil,
lubricants, maintenance materials, miscellaneous support
supply, depot level reparables, and temporary additional duty.
Contract costs for the aircraft comprise what was thought of as
e
oetar reh potential
aneac.Wt
orhLn
hr +-..
Third
and
With
.....
r;n Fourth
te Line maintenance.
wass
at the
s...c.uring of te PA.F, it was advised that m'ight be more
applicable to treat these values as annual contract costs. The
costs can be split among the three main aircraft systems propulsion, and avionics - and supply. All of the
contract cost models need updating, and work is currently
underway to improve accuracy and increase the number of
cost drivers.

Supply contract costs capture the cost of

5-5

shipping airframe, engine, and avionics components between

the base and the contractor, and some of the costs for the
supply of unit level consumption materials. The final two
O&S cost models deal with Sustaining Support, and
Installation Support Funds. Sustaining support includes the
cost of replacement support equipment, modification kit
procurement, and sustaining engineering support. Installation
support costs are made up of personnel pay and allowances,
material, and utilities needed for the maintenance of the base.

EXAMPLE RESULTS - In order to demonstrate the model

capabilities, a number of aircraft solutions to a single mission
specification were generated, optimised for minimum LCC.
Aft-tail, delta, and delta-canard designs were produced with
single and twin engines, and with options for crew and fin
numbers. A notional air-intercept and combat mission was
used, shown in Figure 6, together with rigorous point
performance parameters, to produce a range of conventional
semi-stealthy high-performance combat aircraft.

The sum of all of the previously calculated values gives the

total O&S cost for one year in 'then-year' dollars, which is

then multiplied by the number of years in service, and can be
'discounted' using standard techniques. The model does not

H=91440.M= 1

H=9144m,.M1
0.9.t = 31,,ST.

oe...

currently contain timescale estimates for development and

production,

making

the

discounting

process

H 8382., M=0.90

slightly

inaccurate, as a lack of discounting applied to the acquisition

then-year O&S cost divided by the number of aircraft on the

base. The O&S phase typically contributes about 50% of the
LCC for modem combat aircraft, and it is in this area that the
largest LCC savings can be made. Reduction in O&S costs
can
achieved by
reducing
one of many
contributing
factors,
e.g. be
maintenance
effort,
fuel burned,
aircrew
numbers, etc.

M= 1.4

H9144.M.=O.85

H9

M0

Den-,LLand &Taxi
=9

applied if discounting were ignored. The cost calculated by

the LCC model, for the individual aircraft, contains the total

Supmrome-25bk

phase will result in those costs appearing to be larger than they

really are. The discounted O&S costs for a life of thirty years

at a rate of 6% per annum is found to be only 13.765 times the

first annual cost, as opposed to the 30 times that would be

Comt
Coma

.r3o-,
-5 1-

M= 075.

A-lot
Waooop, Taxi

Disposal - This could amount to a wholly negative cost if the

aircraft were sold intact at the end of its useful life. As this is
very unlikely, the disposal cost model consists of the
following contributors; disassembly labour, disposal of nonreusable material, sale of scrap material, and resale value of
on-board equipment. Depending on the relative values of
these different components, the total disposal figure could be
positive (i.e. a cost), or negative (a credit). The resale of
systems, such as the engines and avionics, is thought to be
unlikely, as technology in these areas changes so quickly that,
for the moment, the value of these items has been neglected.

Figure 6. Aircraft Mission Performance Requirements.

Ground roll was limited to 250m for take-off and 500m for
landing with full mission payload (475kg) of external
lweapons,
and and 95%
95%
fulfuel. . Two
Two attad
turns wereofset;
eternat
attained turns
7g at
12750m, M = 1.4, and 7.5g at 10805m,
M = 0.9. Two
sustained turns with reheat were specified; 5g at 6050m, M =
0.8, and 3g at 12825m, M = 1.6. All turn constraints were
said to have full mission weapons load and 50% fuel
remaining. Excess thrust values were checked at 1i1500m, M
= 2, and sea level, M = 1.2, with full weapons load, and 50%
and 80% fuel respectively. The final point performance
constraint was for a time to climb. Initially, the aircraft was at
sea level and M = 0.25, climbing to 9144m and M = 1.5 in 90
seconds; mission weapons mass and 50% fuel were assumed
at the start.
The aircraft is assumed to be built by a collaborative group of

Total Life Cycle Cost per aircraft is simply the sum of the
different cost phases already calculated, apportioned to
different numbers of aircraft, depending on the Life Cycle
phase. Figure 5 gives an example then-year LCC breakdown
for a modem combat aircraft, having a large composite
materials content, which has the effect of increasing disposal
costs. Encouragingly, O&S costs now appear to make up a
smaller fraction of the total LCC compared with the last
generation of combat aircraft, where operation and support
costs typically contributed 60-70% of total LCC.

two major and two minor partners, with a total buy of 620
aircraft and a production rate of 4.5 per month. FY2000 was
assumed as the accounting year. The aircraft is to be operated
from dispersed main operating bases of three squadrons (39
aircraft) per base, with a three-tier maintenance strategy - First
and Second Line on the base, as well as Third Line contracts.
It is to have a life of 25 years, at 240 flight hours per year,
giving a total flying life of 6000 hours. 'Deep overheads',
such as the cost of fighter control, are ignored, as are other
costs not affected by the design of the aircraft.

Disposal

O&S

0 123

Production

Figure 7. Twin-Engine Configuration Solutions.

GSE&IS
Figure 5. Approximate Then-Year LCC Contributors.

Visualisations of the resulting one-crew, twin-engine, singlefin, configurations are shown in Figure 7. As can be seen, the
delta and delta-canard designs are smaller than the equivalent
aft-tail aircraft, although the delta will have lower agility than

5-6

the other two. The aft-tail aircraft can probably be the most
'stealthy' with the least compromise from near-optimum
placement of control surfaces, which would most significantly
affect the delta-canard. Thus, all of the proposed solutions are
viable designs having particular strengths and weaknesses.
Fuel, empty, mission, and gross masses are shown in Figure 8,
where mission mass is the mass of the aircraft, including crew,
fuel, and weapons payload. The mass figures confirm the
relative sizes of the aircraft, with the delta and canard-delta
aircraft seen to be the lightest.

A9-7.j I IC-,

C
,

C...d I C-,,
F..

99.0
1

2 C,.,,d 2CF
.

00

14564
20999

M62
241,7

6....

Cn
I C..d
Fj

47,

213

M
Mus

664,,,, 6..,
M_
3

I W1. 'IC.
P6,
F,,

5473

F.I M_

15033

I C..orn1
OC-w2
Fl,

It can be seen that for every equivalent aircraft configuration

(except the single engine aft-tail aircraft, which is not a fully
converged solution), the mass of the twin engine aircraft is
lower, but the LCC has increased. This result is interesting for
two reasons. Firstly, it questions a commonly held perception
that, for a given set of requirements, a single-engine aircraft
will
be smaller
and that
lighter
a twin-engine
configuration.
Secondly,
it shows
for than
significant
configuration
changes
shorthat fo ss.
it
cost
cost is not proportional to mass.
This second observation prompted further research in to the
differences between aircraft optimised for mass and LCC. For
both single and twin-engine canard configurations (which had
previously been judged to offer a good balance of cost and
"combat effectiveness), solutions were produced with gross
mass as the optimiser objective function. The results of the
comparison appear in Figure 11, where the relevant objective
function is minimised in all cases.

U
W

530
' 41

062

16127
21

16773
29

11470
44
24

,7
j474

Figure 8. Twin-Engine Mass Breakdown.

The costs for the various configuration options are shown in

Figure 9. As expected for aircraft of similar technology levels,
the costs change roughly in proportion to the mass, with all

"

proportionate mass increases being greater than the relevant

cost increases, except for the delta-canard design. This is due

to a change in the driving performance constraints between the
delta and canard-delta solutions, resulting in only a very small
mass increase, but a larger cost increase, due to the increased
complexity of the canard-delta aircraft.

Afl-TUt
I~n
Cir- I
Fn

D,t Im
Cr' I

Cmlad IlCmw I

F.,

F,,

CmludinICre 2 Cmawd
2 C"
Fi.

I~iCmwad2 C 2t

Fo,

0.9n

o.-6

0.798

o.99

04.83

0484

O&S

9994

9247

92.662

95.122

99.058

100.812

6243

5.436

3.&43

1.

5.872

,3.%

,.,

Figure 9. Twin-Engine LCC Breakdown.

A comparison of the relative mass and cost increments relative
to the single-engine delta (the lightest and cheapest aircraft) is
presented in Figure 10.

ogroup.

1LC-S-i ,
M1CC -Too

13

Fn

Fl,

Fm I

Fi

000

.022

.07

.076

1.0

1.257
.1.179

2oiE

Singl.E,,n,

T.

I.CC-Optbirr

m.9
09.o57

,Eni.

_aOptimi

o,0

T mFon
LCC-Optlmid

o.C,
,3

Figure 11. Relative Mass & LCC - Mass vs. LCC-Optimised.

Although the differences are small when presented as above,
the total savings for an aircraft programme, with no loss of
capability, are significant. The savings above correspond to a
monetary value of FYOO$M970 for the single-engine aircraft,
and to FYOO$M435 for the twin-engine solution. These
savings increase once discounting is included, due to the
reduced influence of the slightly increased O&S costs. In
future, the difference between aircraft optimised for mass and
cost will increase as the R&M and LCC models are improved
to reduce their dependency on mass.
effects of internal weapons carriage, composite materials
usage, and more electric aircraft (MEA) technologies on
aircraft mass, performance, and cost. The LCC model requires
further development, particularly for the R&M model, and the
manpower scaling from it. This will improve accuracy and
confidence in the models, and should produce greater
differences in aircraft optimised for mass and LCC. Longterm, this will allow greater savings in through-life cost, as
aircraft become closer to true minimum-cost designs. What
the objective function should be, and .the confidence that is
placed in it are subject to discussion; increased modelling and
understanding of the real cost drivers can only be of benefit.

improved propulsion calculations, and has been validated

...

0611.

Future developments - New versions of the cost model are

intended to be included with the more detailed design and
optimisation tool used by DERA's Air Vehicle Performance
This model offers higher aircraft modelling fidelity,

07

M.

Smgk
o.

M-OCtmid

Fi.

UGSES

3.352

.4

1.058

14
L.07

1.44
t

.005

F-

.073
1.123

.100
I

Figure 10. Relative Mass and LCC - Single vs. Twin Engine.

more thoroughly than the current aircraft synthesis model.

The combination of the improved LCC model, higher fidelity
aircraft models, and up-rated optimisation software should
dramatically increase DERA's capability to 'Design for LCC'.

This should enable significant improvements in the costeffectiveness of future British/European combat aircraft.

5-7

PART II - THE LOW SUPPORT VEHICLE

Mission:

hi-lo-lo-hi penetration

OVERVIEW - A fresh approach to solving the problem of

affordability of future combat aircraft was proposed[9],
involving the formulation of a Low Support Vehicle (LSV)
concept, specifically directed at minimising through-life
operating and support requirements. The outline specification
for the concept design of the LSV, and a description of the
principal characteristics of the LSV, was established[ 10], so
that a single detailed concept formulation could be produced.
The Department of Aeronautics, Imperial College, London
undertook the study, under a three-year contract. The purpose
of the study project was to demonstrate the effects of aiming
for drastic reductions in support costs for all parts of a combat
aircraft design, so allowing a better trade-off to be made
between support characteristics and other features.

Payload weight with max. fuel:

4000 kg

Max. low-level speed with stores:

Maximum load factor:

Mach 0.92
7.5g

This section of the paper summarises the evolution of the LSV

design and indicates some of the supportability features
incorporated. The method developed for the prediction of the
reliability and maintainability (R&M) of the LSV is described,
together with the outcome of the predictions. A separate
assessment of the cost-effectiveness, design attributes, and
supportability features of the LSV has been conducted, but is
outside the scope of this paper.
Supportability - For the purpose of the study, a supportable
aircraft was defined as one with low support requirements; a
minimum expenditure of equipment, effort, and, ultimately,
money is required for the aircraft to fulfil its assigned role.
Supportability of the aircraft alone was considered as a
fundamental consequence of R&M. This was derived from
the assumption that everything that affects the direct support
of a system, other than consumables, can be linked to the
inherent R&M characteristics of the system. For the study
therefore, supportability is the quality possessed by a
supportable aircraft.
The usual perception of supportability is that it is simply the
consequence of the R&M characteristics of the aircraft under
consideration. However, this view can be broadened to
encompass the supporting systems required by the aircraft to
complete its mission. This brings in such systems as in-flight
refuelling tanker aircraft, or escorting fighter aircraft. This
idea of total
reflects
aircraft's
capability
as system
well as support
its R&Mrequirement
characteristics,
and the
is dependent
upon the aircraft's mission,
It was considered that a total system support approach would
be used to evaluate the supportability of the LSV in
comparison with other combat aircraft. A framework for the
method of comparison was developed during the study, which
could be completed with the application of the appropriate
analysis tools. The concept formulation for the LSV took into
account the direct influence of R&M. It also considered the
impact on the total system supportability in terms such as the
ability to operate from short or damaged runways, the need for
air-to-air
deployability of the aircraft by
provision refuelling,
of systemsandthatthewould
reduce the need for an
extensive logistics tail.
Evolution of the LSV design - The LSV was intended as an
exploration of the effects of designing to minimise the aircraft
support requirement. The specifications[10] called for the
design to be formulated for the offensive role, with key
performance specifications being equivalent to the Tornado
GR4. Deviations from the performance requirements were
only to be made to reduce the support requirement without
significant reduction in capability. The essential features of
the Tornado GR4 adopted for the LSV are shown below:

The LSV was initially specified to have a hi-hi mission radius

of 1400nm, compared to the published Tornado combat radius
(un-refuelled) of 750nm, although this was subsequently
modified to a more representative hi-lo-lo-hi profile of
1130nm. Take off and landing distance of 1565m was
specified on the grounds that a supportable and deployable
concept such as the LSV should be capable of operating from
many different bases - published data shows 36 British civil
airfields with runways of adequate length for this
performance. Such performance is comparable to that of
modem mid-sized civil aircraft such as the Boeing 757 and
Airbus A319, indicating that a similar distribution of suitable
airfields should be found elsewhere in the world. Although an
unclassified study, the LSV project took into account open
literature information on low observables and considered them
as part of the design.
DESIGN PROCESS - It was appreciated that the LSV would
be a novel configuration in many respects, so that it was not
appropriate to design just a single aircraft using existing
methods; several different solutions might be possible for
some areas of the design. However, without actually reaching
the final design process for the aircraft it was not possible to
identify which features should be included, and where new
methods would be required. To resolve these problems, a
scheme was established for developing several configurations
leading to the ultimate design. New methods were developed
in parallel with the development of the configurations, so that
the final configuration would incorporate not only the best
design features, but also the most refined calculations.
T d
The design philosophy of the LSV was intended to maximise
supportability, but the approach leads to some significant
impact on the overall design. Inherent component reliability,
although obviously desirable, cannot ensure reliability of a
complex system due to the large number of individual
components.
the operating
environment
for subsystems and Improving
even systems
can significantly
improve
the
reliability of an aircraft, but such improvements can only be
achieved if they are considered early in the design
process.
For example, providing a better environmental control system
and planning the layout of all systems to provide a favourable
operating environment can only be done by giving
supportability a high priority early in the design process.
Similarly, de-rating systems, particularly the engines, may
provide a more benign environment, thus improving
reliability. However, to achieve the same performance targets,
de-rated engines must be larger than those operating at their
maximum rating, affecting much of the design.
The LSV philosophy also stresses simplicity and integration as
means to improve reliability. The use of integrated avionics
systems, capable of re-configuring to take over functions
of
failed units is one example, but the concept of integration and
simplicity can be applied to a much more basic level. For
example, the weapons bay door on the LSV combines many
functions, facilitating the release of internally carried
weapons, and providing access to refuelling points and other
internal systems. The complexity of the weapon bay doors is
not increased, but the increased functionality eliminates the
need for extra doors for refuelling and maintenance.
Accessibility of systems for maintenance is emphasised in the

5-8
early design phases; an inadequate initial basic configuration
and layout of systems can never be recovered in the detailed
design phase. Finally, the LSV design philosophy avoided
reliance on the use of unproven or speculative technologies,
the failure of which to realise their potential would
fundamentally undermine the ability to achieve the LSV aim.
From the target specifications, a baseline configuration was
formulated. The designations LSV A and LSV B were used
during the baseline formulation so that the first configuration
to be designed was called the LSV C. The LSV C is a single
seat, flying wing aircraft with wing-tip fins and a short nose.
It is predominantly

constructed

of composite

As a result of the experience gained during the design of the

LSV C, in both configuration and methodology, two further
LSV configurations, the LSV D and the LSV E, were
produced. The LSV D is a low aspect ratio delta wing design
with no horizontal or vertical tail and no protruding nose. A
single chin engine intake leads to the single non-afterburning
turbofan engine. The LSV E has a planform similar to that of
the LSV C, but without the wing tip fins, being a flying wing
with no protruding nose. The trailing edge has a pronounced
kink around 40% of the semi-span. Two engine inlets are
positioned on the upper surface of the wing, each feeding a
non-afterburning turbofan engine.
Analysis of the LSV-D and LSV-E showed that both met their
targets, and could be developed into extremely supportable
aircraft. The LSV-E was considered a more operationally
flexible design, so this configuration was chosen for
optimisation, to produce the final LSV configuration, the
LSV-F. The computerised methods used to design the LSV C,
LSV-D and LSV-E were combined in an automated design
synthesis very similar to the one described in Part I of this
paper. Many of the aircraft design models, particularly mass
and aerodynamic estimation methodologies, had to be updated
to allow for the unconventional design, although the
thermodynamic engine models were derived from the same
source. Figure 12 shows the optimiser evolution of the
configuration, with drawings of the aircraft at the start, after
14 and 28 iterations, and at the end of 42 iterations,
533%

.....
....
..
............

......

_..

materials.

Internal weapons bays flank the single non-afterbuming, derated engine. The LSV C was developed from the baseline
configuration, and has a 500 leading edge sweep with leadingedge root extensions. The trailing edge is kinked, and
supports four control surfaces - a pair of rudders and a pair of
elevons. The aircraft is slightly statically unstable.

Ie t~mU.sr 0%

--

,
Figure 13. Final LSV-F Configuration.

Figure 14. LSV-F in Flight.

LSV-F SUPPORTABILITY - The LSV concept design is
driven by the need to reduce support requirements. Aircraft
supportability is achieved mainly by simplification, systems
integration, redundancy, and by commonality of parts. The
need for supporting systems is reduced by making the aircraft
capable of autonomous operation; long range without
refuelling, low-observable features, comprehensive electronic
counter-measures, high speed at low level, and self-defence
weapons all contribute to this. For maintenance purposes, all
important systems can be accessed from the ground, without
downloading weapons, via either the avionics bay, the cockpit,
the undercarriage bays, or the weapons bay. On-board oxygen
and inert gas generators (OBOGS/OBIGGS) and a multifunction integrated power unit reduce the need for ground
support equipment.

10M7

Be

The LSV-F structure employs composite materials to reduce

Eeptywtght: 11,997kg

10,025k9g

Tokof Woght 29.645 kg

25.149 kg

A
a 0

fatigue, corrosion, and weight. A modular structure is used,

possibly incorporating a damage sensing system, to allow oncondition maintenance and to reduce peacetime operating
costs, through reduction of third and fourth line maintenance
and no-fault-found (NFF) reporting.
Many structural
components are common, and there are no leading and trailing
-

edge high-lift devices. The engines are de-rated, and do not

kg

have reheat capability. Engine installation and removal is

achieved by means of a special trolley, and all maintenance
actions can be effected from below, so avoiding damage to the

.755 k
24.347 kg

9,8 kg
24,139

Figure 12. Optimiser Evolution of the LSV-F.

The LSV-F is a twin-engine flying wing, of aimost pure deita
planform, with a trailing edge kink at just under 50% semispan. Two split elevons are the only control surfaces, and the
aircraft is slightly unstable in pitch. It has a single centreline
weapons bay with two doors, as well as the provision to carry
external stores. The LSV-F configuration is illustrated in
Figure 13, and by a computer generated image in Figure 14.

upper wing skin. Due to the position of the intakes, there is

little danger of foreign object damage (FOD) to the engines.
The main undercarriage is very simple and robust, the units
being interchangeable between the left and right sides of the
aircraft. Oversized tyres operating at low pressure give
increased tyre life, and two wheels per main undercarriage
strut reduce the kinetic energy per wheel, allowing simpler
brakes. The nosewheel uses the same type of tyre as the

5-9
mainwheels, and the oleo shock absorbers are identical for the
main and nose undercarriage. A titanium matrix composite
could be used for the undercarriage, eliminating the need for
corrosion inspection.
Whole fuel tanks are formed from composite material, to
reduce leakage at tank joins. The tanks are foam-filled, and
can be pressurised from the inert gas generating system. The
weapons bay provides a benign environment, improving the
reliability of weapons that may be carried on a number of
missions without being expended. The gun is positioned to
prevent interference with other aircraft systems and minimise
the effects of vibrations from gun firing and ingestion of gun
gas by the engines. A disposable cover is fitted over the gun
port, and ammunition replenishment is carried out via the
starboard main undercarriage bay. The avionics bays are
easily accessible, and the windshield may be opened to access
cockpit avionics. All sensors are readily accessible from
ground level. Avionics reliability is enhanced by a closed
loop environmental control system for the avionics bays.
The hydraulic system is of simple configuration, employing
electro-hydrostatic actuators. The ultimate goal is to eliminate
sppor cots.The n-bard
hydraulicsstudy, hydrulis tofurherredue
multifunction power unit provides engine start, emergency
power, and auxiliary ground power. The power unit replaces
ground support equipment, as does the on-board inert gas
generator. The main utility locations are positioned to prevent
any compromise to system accessibility if more than one
maintenance task is being carried out simultaneously. Except
for the cockpit (which has its own access ladder) and upper
wing surface, all points can be reached from the ground
without ladders or stands.
The supportability measures identified for the LSV-F are
summarised in Table 1.

Structure

RELIABILITY
AND MAINTAINABILITY
Two
quantitative measures of aircraft reliability and maintainability
are generally available. These are the defect rate (DR),
usually expressed as defect occurrences per 1000 flying hours,
and the defect man-hour rate (DMHR), which is the number of
man-hours
defects,inagain
expressed
per
1000 flying spent
hours.rectifying
The termthe'defect',
the context
of this
or adfaut, reuiin contectiv
fers to ure
stud
corrective
fault, requiring
failure or
to a (referred
refers action
maintenance
to asa 'rectifying'
the fault).
The defect rate is a measure of reliability; more reliable
aircraft will have a lower DR. The DMHR is often described
as a measure of maintainability. However, the DMHR cannot
be taken as an independent measure of maintainability, since
the man-hours spent rectifying the defects in a given number
of flying hours will depend not only on how easy it is to repair
the system, but how many times it needs to be repaired. An
independent measure of maintainability is found by dividing
the DMHR by the DR, to give the mean time to repair each
defect (MITTR), in man-hours per defect. If the reliability
(DR) and maintainability (MTIR) are known, the unscheduled

The approach of separating reliability and maintainability

measures and then multiplying to find the overall maintenance
requirement, rather than attempting to predict the DMHR
directly, was considered to lead to a more accurate and robust
prediction method. The two aspects of the OMR are driven by

full-length engine access doors

simple engine removal concept

corrosion-resistant undercarriage
low pressure tyres
simple brakes
integrated avionics and sensors
low number of hydraulic system functions
replacement of secondary hydraulic actuators
with electrical systems
multi-function integrated power unit
on-board oxygen generating system
on-board inert gas generating system

Maintenance

Table 1. Supportability features of LSV-F.

composite fatigue-resistant-airframe
only four multifunction control surfaces

interchangeable undercarriage components

wheels
single tyre and oleo types for all

Systems

single crew
in-flight refuelling if required
in-lght relling rqe if
integrated weapons loading/launching arm
internal weapons carriage
self-defence capability
long range with internal fuel

overall maintenance requirement (OMR) can be found.

fixed geometry air intakes

reduced likelihood of FOD to engine

Alighting

Operational

simple modular construction

non-afterburning, de-rated engines

Propulsion

capable of operation from civil airfields

stealth and advanced electronic countermeasures allow operation with minimum
support

very simple configuration

access-driven design
accessible avionics bays, >50% growth space
easy radar access

different factors. Reliability is dependent on factors such as

complexity, loading, and component reliability, whereas
independent features such as accessibility and test methods
determine maintainability. The three measures of merit (DR,
DMHR, and MTIR) were used as part of the overall LSV
supportability assessment method, the most important part of
which was the prediction of the LSV R&M.
Different R&M analysis and prediction methods are
appropriate for different stages of an aircraft design process.
Several existing reliability and maintainability prediction
methods were examined during the study to determine their
applicability to the LSV design process. It was concluded that
insufficient data was available to use detailed design methods
for the supportability analysis of the LSV. Further, existing
methods for conceptual design analysis were too old and too
simplistic. It was therefore necessary to develop a new
method for the prediction of the LSV reliability and
maintainability for use in the supportability assessment.
R&M Prediction - For the purposes of the study, aircraft were
considered to consist of twelve systems: air conditioning;
flying/operational controls; fuel system; hydraulic power and
pneumatics; alighting/arrestor gear; oxygen; miscellaneous
utilities; structure system; propulsion systems; armament
systems/tactical avionics; navigation and communications
systems; electrical and instrument systems.

5-10
The R&M prediction method used consists of a set of
statistically derived equations, based on work by Harmon[ I1],
and updated by Serghides[12].
The equations predict,
separately, the reliability and maintainability of aircraft
systems, which can then be combined to give total aircraft
figures and an overall figure for the support requirement in
man-hours per flying hour. Data for ten aircraft in current or
past RAF service, plus two US-operated aircraft (used to
derive only the reliability equations) were collated, mainly
from official sources. All aircraft are jet-powered combat
aircraft from advanced trainers through interceptors and strike
aircraft to a long-range strategic bomber, although the results
should be treated with caution for such a large aircraft.

Only the results for Lightning and Gnat are poor predictions;
these were both significant outliers in some of the derivations
and were excluded from the derived equations. The policy of
excluding outliers tends to increase the total error of the
excluded aircraft, but the resulting equations better represent
engineering trends, and thus should have superior predictive
ability. Plotting of actual and predicted rates for each aircraft
by system (see Figure 16 for the Harrier GR5/7 as an example)
indicate that the accuracy of the prediction is a consequence of
good system level prediction, rather than fortuitous
cancellation of system errors.

The purpose of the prediction equations is to relate measurable

physical parameters describing the aircraft to observed

i
2

measures of R&M. The accuracy of the method depends very

much on the consistency of the data. In addition, there are
other factors affecting R&M that will not be accounted for in
the equations, but which could cause errors. Such error
sources include the data collection procedure, definition and
capture of variables and data, aircraft reliability growth,
inconsistent maintenance policies, differing operating and
environmental conditions, small sample size, and the use of
few parameters to reflect complex design effects. It has been
assumed that the influence of these factors is relatively small.

.
_

Figure 16. Predicted and Recorded Defect Rates, By System.

Maintainability prediction eguations - As noted above, the
maintainability equations were derived using data for the ten

Derivation of prediction equations - Equations for all twelve

aircraft systems were derived using multiple regression
analysis of the defect rate and mean time to repair data with
various parameters, alone and in combination with others.
Over 90 parameters were tested, of which 28 and 19 were
finally selected for the reliability and maintainability equations
respectively.
Several possible forms of equation were
investigated in each case, some incorporating the influence of
time on reliability (called time improvement factor, TIF)
resulting from different technology standards and design
practices. Prediction
equations
only accepted if
there
was oun
engneeing
asi were
fo theincusio
ofthe
was sound engineering basis for the inclusion of the
parameters, and the trends produced were logical. Points that
did not fit an otherwise
clearly or
defined
trend, either
as a result
of kownexcptinalfeaure
s areslt
f aknon
eror
of known exceptional features or as a result of a known error
source (known as outliers), were discarded. For example,
Lightning landing gear data was discounted, as the aircraft
was notorious for its poorly sized tyres, which were very
narrow in order to stow within the thin wings of the type.

aircraft in the database operated by the RAF. Maintainability

equations for the MTTR of each of the twelve aircraft systems
were derived using the same multivariate regression process as
used for the reliability predictions. Time improvement factors
are not used in the maintainability prediction equations, since
maintainability is not dependent on component design to the
same extent as reliability. However, factors are employed to
account for 'design for maintainability', and the resulting
improvement in the accessibility of systems.

Total aircraft defect rates - The total aircraft defect rate is

found by adding the system defect rates. The accuracy of the
total defect rate prediction is well illustrated by using the
reliability equations to compare the predicted and recorded
reliability of a number of aircraft. The results of the
comparison, together with a best-fit line and 'goodness of fit'
metrics are shown in Figure 15.

The
total aircraft
mean
to repair
is defined
as therate.
totalThe
of
the defect
man-hour
ratetime
divided
by the
total defect
reutfomheRMpdiineqaosmybec
bnd
results from the R&M prediction equations may be combined
tflying
oghour).
rat angove
rt of
(man-hour
per
Figure 17 defect
shows a man-hou
comparison
predicted and
recorded defect man-hour rates for various aircraft.
-

y. 1.....

....

_____
_

____

"
_

A
.
,

I, IMUo12
-R2. .0-72

-LSV-F

...
I
T......
_

-.

~....
.upon

...

___

_-

_
_

'.

___

~to

dDd ft,0

Figure 15. Predicted vs. Recorded Defect Rates.

-- i:

o.944i

-,,R.,ordod

-~

NetecM-KMlln-HoorR.I
R.0e

Figure 17. Predicted vs. Recorded Defect Man-Hour Rates.

PREDICTIONS - The R&M forecasts for the LSV-F
were produced using the prediction equations, and drawing
47 available inputs. The inputs ranged from the date of
the first flight of the aircraft type, through empty weight, to
whether the aircraft carried a primary radar. It should be
noted that the outputs from some of the prediction equations
were modified to account for special features of the LSV-F,
such as additional effort expected in the detailed design stage
reduce support requirements. If no adjustments to the
predictions are allowed, the DMHR is increased by 9%.

5-11
It would be misleading to compare the predicted defect manhour rate and its components, the defect rate and the mean
time to repair a defect, for the LSV-F with published data
available for other aircraft. The LSV-F predictions are based
on the maintenance policies and data collection standards of
the RAF. Other aircraft operators use figures derived from a
different base; some include planned as well as corrective
maintenance, or consider on-aircraft maintenance time only.
The study compared the predicted rates for the LSV-F with the
predicted rate of other combat aircraft. Figure 18 shows the
results of the comparison, scaled relative to the predicted
values of Tornado GRI.
.

,............. 1of

S-.Vehicle,

basis of an updated model, which will be compatible with the

LCC model described in Part I. This will contribute to a
powerful and flexible suite of aircraft design and analysis
tools, capable of designing and optimising for either minimum
mass, LCC, support effort, or eventually, availability-cost.

CONCLUSIONS
This paper has described the methods and results from two
DERA-sponsored University research programmes. The first,
performed by the College of Aeronautics, Cranfield
University, developed a computerised design and optimisation
tool to minimise the Life Cycle Cost of combat aircraft. The
tool and the results from some studies were presented in Part I
this paper. The second research programme, performed by
the Department of Aeronautics, Imperial College, produced a
similar tool to investigate the supportability gains that could
be achieved by an aircraft designed for maximum
supportability.
The resulting aircraft, the Low Support
and the methods used to assess its supportability
were described in Part II of this paper.
Both research programmes have shown that design for
reduced cost is possible, but that quantifying the benefits is

S
Figure 18. Predicted R&M Quantities.
The comparison shows that the defect-related maintenance
requirement (the DMHR) of the LSV-F is considerably lower
than that of all the other aircraft, being nearly half of that
predicted for Eurofighter and nearly a third of that predicted
for Tornado GRI. The maintainability prediction for the
Lf - structure
sTructuGr1.e
pe
intnabiseslthetuse
mpodicteand
LSV-F
penalises
the
use of composites
and the
relatively poor accessibility due to the low aspect ratio allwing configuration. Except for the structure system, the
mainteconfigurance r iEment
.
every
fof
tructr
systemf
Lpeace
the
maintenance requirement of every system of the LSV-F is less
than that of each of the other aircraft. The relative advantage
of the LSV-F over the other aircraft is less pronounced in
maintainability (the MTTR) than in reliability (the DR). It is
likely that some maintainability advantages of the LSV-F are
not reflected in the maintainability prediction equations.
Influence of technolony - The LSV-F (and F-22 and
Eurofghter) predictions benefit from the increased reliabilityof
of modem systems reflected in the Time Improvement Factors
in the reliability prediction equations. By eliminating this
factor it is possible to judge whether the LSV-F and its
'contemporaries' are the most supportable simply because
they use newer technology, or whether the designs are
fundamentally superior. Setting the technology datum to
2002,dthentalsumedfiorst
2002, the assumed first posslettinghte
possible flight dtecnogy
date of tsumlts
the LSV, results
in an overall flattening of the distribution of the results,
although the pattern is very similar. The LSV-F remains best
by a considerable margin, but its nearest competitor becomes
the Jaguar, a far less capable aircraft. F-22 slips to fourth in
the ranking, being overtaken by the less capable F-16. The
implication of this is that the supportability advantage of the
LSV-F derives partly from the application of new technology,
but mainly from the fundamentals of the design.
Future Developments - As the LSV-F was single point design
for a given mission specification, it is unlikely that the actual
design will progress further than at present. However, the
design philosophy and many of the models will be taken
forward and updated for use with other DERA aircraft design
and analysis tools. Of particular interest are the mass and
aerodynamic estimation methods for the flying wing
configuration. The R&M prediction equations will form the

difficult and requires extensive modelling effort. The LCC

model has shown that reduced through-life cost will not
always be achieved by reducing support costs. Although O&S
contributes approximately 50% of the through-life cost, the
economic impact of increasing reliability and compromises to
the design may outweigh, in life cycle terms, the benefits of
reduced O&S costs. This matter is further complicated by the
difference between discounted and non-discounted costs.
The design characteristics resulting from the cost and support
design drivers are considered to improve future aircraft
supportability, and therefore improve future combat aircraft
time and war time availability, whilst reducing throughlife costs. This in turn should lead to aircraft capable of
delivering a set level of performance for reduced cost,
m x m sn
vl e
t e military
mlt r sense,
e s ,aandd leading
l a i g tot a
maximising 'value' inn the
situation that will be beneficial for both customer and
manufacturer alike.
Acknowledgement - Acknowledgement is due to Andy
Clifford of the Centre for Defence Analysis, DERA, for
much of the material in Part of this paper.
REFERENCES
1
2
2
3
4

Woodford, S (1999).
The Minimisation of Combat
Aircraft Life Cycle Cost through Conceptual Design
Optimisation. PhD Thesis, Cranfield University.
Whittle, R
G Thesis,
(1998). Imperial
The Design
of the
Low Support
Vehicle.
PhD
College,
London.
Mattingly, J D (1996).
Elements of Gas Turbine
Propulsion. McGraw-Hill Inc.
Bums, J W (1994). Aircraft Cost Estimation and Value
of a Pound Derivation for Preliminary Design
Development Applications. SAWE Paper 2228.
Birkler, J L, Garfinkle, J B, and Marks K E (1982).
Development
and
Production
Cost
Estimating
Relationships for Aircraft Turbine Engines. RAND
Corporation N- 1882-AF.
Stemberger, N L, et al. (1980). Modular Life Cycle Cost
Model for Advanced Aircraft Systems, Phase III.
AFFDL-TR-78-40. USAF Flight Dynamics Laboratory,
Wright-Patterson Air Force Base.
Anon. (1992). Operating and Support Cost-Estimating
Guide. Office of the Secretary of Defence - Cost
Analysis and Improvement Group.
The Pentagon,
Washington DC.

5-12

Anon. (1980). Naval Aircraft Operating and Support

Cost-Estimating Model
FY-1979.
ASC-R-126.
Administrative Sciences Corporation, Falls Church, VA.
9
Gill, A D (1992). A methodology for reducing aircraft
life cycle cost. DRA Working Paper. OS34-92-WP-18.
10 Gill, A D (1993).
Specification and principal
characteristics of the low support vehicle. DRA Working
Paper. OS34-93-WP-17.
11 Harmon, D F, Pates, P A, and Gregor, D (1975).
Maintainability Estimating Relationships.
IN:
Proceedings of the 1975 Annual Reliability and
Maintainability Symposium.
12 Serghides, V C (1985). Development of a Reliability and
Maintainability Prediction Methodology for the Aircraft
Conceptual Design Process. M.Sc. Thesis, Cranfield
University.
British Crown Copyright 1999 /DERA
Published with the permission of the Controller of Her
Britannic Majesty's Stationery Office.